More than half the prescription drugs approved by the Food and Drug Administration (FDA) between 1976 and 1985 caused serious side effects that later caused the drugs to be either relabeled or removed from the market, according to a General Accounting Office (GAO) study released yesterday.

The side effects included reactions that could cause hospitalization, disability and even death.

The risks posed require that the FDA improve its monitoring of the safety of recently approved drugs, the congressional watchdog agency said.

But the GAO acknowledged the study did not determine why the side effects occurred, whether they could have been identified by the FDA before approval and how many patients were affected. The report also prompted an unusually blunt response from the Department of Health and Human Services, which called it methodologically unsound, alarmist, and misleading in the attention it gave to problems that were largely unavoidable.

"The draft report does not discuss the benefits to be gained by the use of drugs, including those with severe risks," HHS Inspector General Richard P. Kusserow said in a lengthy rebuttal to the document. "Some of the drugs that have had post-approval identification of risks are also the only available treatment, the best available treatment, or second-line treatment for use when other drugs have failed. By not looking at both benefits and risks -- as FDA must do -- the report misleads readers to the conclusion that the drug review process is inadequate."

The focus of the report was the surveillance system the FDA uses to monitor the safety of drugs after they have been approved. The system is necessary because the human tests conducted to assess a drug's safety before it is approved for sale cannot pick up all the potential problems with that drug therapy.

For example, pre-market drug tests include an average of 3,000 patients over a period of several months or even years. But if a drug has a side effect that occurs only once in every 10,000 patients or only after long-term use or only when the therapy is taken in combination with another, it is likely to be picked up only after the drug is being used by the public.

As a result, in many cases the FDA amends the labeling instructions or revokes a drug's approval within the first few months or years after a medication is approved.

The GAO report found that these side effects were common -- from a wide range of drugs for treating almost everything from infertility to heart disease -- and fairly serious. They resulted in hospitalization, permanent disability and even death.

The report also found that drugs with serious post-approval side effects had a shorter approval time than drugs without such problems, and that drugs approved for chidren were twice as likely to have serious post-approval risks than other medications.

"This is an important reminder that FDA approval does not guarantee that approved drugs are completely safe," said Rep. Ted Weiss (D-N.Y.), chairman of the House Government Operations subcommittee on human resources and intergovernmental relations that commissioned the study.

GAO concluded that the FDA should "establish formal procedures" to evaluate post-approval risks, which could then be used to implement "better and more timely drug labeling."

But in its response, HHS said the GAO did not make clear how the FDA could be expected to anticipate problems that cannot be detected by pre-approval tests and failed to consider the balancing of risks and benefits that the agency performs before approving the drugs.

"The recommendation presupposes that FDA can, in some undefined, vague way, anticipate the unknown and make decisions about the approvability of drugs based on such anticipatory judgments," Kusserow wrote. "We do not believe this is either possible or desirable."